JP5217875B2 - Sound field support device, sound field support method and program - Google Patents

Description

  The present invention relates to a technique for controlling acoustic characteristics of an acoustic space.

There is a sound field support system based on an existing acoustic environment in an acoustic space, and enhancing and correcting reflected sound characteristics including reverberation effects and initial reflected sound in the acoustic space to control the acoustic characteristics. This sound field support system includes a microphone and a speaker fixed to a ceiling or a side wall of an acoustic space, and a sound field support device connected thereto.
When a collected sound signal is input from a microphone, the sound field supporting device of this type of sound field supporting system is for applying a reverberation effect of a desired acoustic space to the collected signal by an FIR (Finite Impulse Response) filter. The reverberation effect in the acoustic space is enhanced by repeating the process of convolving the filter coefficient sequence and emitting the signal obtained as a result from the speaker and returning a part of the emitted sound to the microphone. Depending on the setting of the filter coefficient sequence used for the FIR filter, it is possible to create a reverberation effect as if it is playing in a large acoustic space such as a concert hall, although it is a narrow acoustic space.
By the way, in this kind of sound field support system, as shown in FIG. 9, sound circulates in a closed loop of acoustic space → microphone → sound field support device → speaker → acoustic space. If a sharp peak appears in the amplitude characteristic of the closed-loop transfer function (frequency response), problems such as coloration may occur. Patent Document 1 is an example of a document that discloses a mechanism for suppressing the occurrence of a sharp peak in the amplitude characteristic of a closed-loop transfer characteristic (frequency response) including such an acoustic space. The sound field support device disclosed in the same document measures an open loop transfer function that is a transfer function of the acoustic space itself in a state before the sound is fed back via the sound field support device. Find the inverse characteristics of the transfer function. Then, the impulse response having the reverse characteristic is sampled to obtain a filter coefficient sequence used for the FIR filter.
Japanese Patent Application Laid-Open No. 07-240993

However, as to whether or not a truly appropriate filter coefficient sequence is set, whether or not the user of the sound field support device (for example, an acoustic engineer) has a hearing problem, There was a problem that it was necessary to listen to and confirm. This is because, when a pulse-like sound is input, depending on the setting contents of the filter coefficient sequence, an unnatural sound separated from the flip is heard due to the characteristic that the FIR tap is separated in the sound reproduced from the speaker. There is a case. Moreover, depending on the magnitude of the tap amplitude and the density of taps, there may be a problem that the sound reproduced from the speaker may be heard as a harsh and distorted sound.
The present invention has been made in view of the above problems, and in the sound field support system, not only the occurrence of problems such as howling and sound coloration, but also acoustic engineers and the like actually listen to the reproduced sound. It is an object of the present invention to provide a technique that makes it possible to improve the sound quality of reproduced sound from a speaker without confirmation.

  In order to solve the above problems, the present invention acquires first data representing an attenuation curve of an impulse response in an acoustic space, and collects a speaker disposed in the acoustic space and a sound output from the speaker. Each of the plurality of types of filter coefficient sequences is selected from an FIR library in which a plurality of types of filter coefficient sequences set in an FIR filter connected to a sounding microphone to form a closed loop is stored. Attenuation curve data acquisition means for acquiring second data representing an attenuation curve of the first and a peak of an attenuation curve represented by the second data and an attenuation curve represented by the first data for each of the second data When the positions are drawn in an overlapping manner, the portion of each of the portions of the former attenuation curve protruding from the latter attenuation curve has its portion A first index obtained by multiplying a weight corresponding to a position on the time axis, and a second index indicating the similarity between the attenuation curve represented by the first data and the first index. Index calculation means for calculating from second data, and among the plurality of types of filter coefficient sequences, the first index is a value in the vicinity of a predetermined first threshold, and the second index And a filter coefficient sequence setting unit that selects and sets the FIR filter that has a similarity equal to or greater than a predetermined second threshold, and is executed by each of the units And a program for causing a computer device to function as each of the above-described means.

According to such a sound field support device, a sound field support method, and a program, a plurality of types of filter coefficient sequences stored in the FIR library are included in the FIR filter that forms a closed loop together with the microphone and the speaker arranged in the acoustic space. Are set to satisfy both of the following conditions (A) and (B).
Condition (A) A portion of the former attenuation curve protruding from the latter attenuation curve when the attenuation curve of the impulse response of the filter coefficient sequence and the attenuation curve of the impulse response of the acoustic space are drawn to overlap each other. The first index obtained by multiplying the degree of the protrusion in each of them by the weight according to the position on the time axis occupied by the portion is a value in the vicinity of a predetermined first threshold value.
Condition (B) The similarity between both attenuation curves is equal to or greater than a predetermined second threshold.
Here, it is conceivable to use various types of the first and second indicators, and what values are suitable as the first and second threshold values depending on the types of the indicators. Is determined.

ただし、数１のΣ演算は、以下の数２を満たす時刻ｔに関してのみ実行し、数１にてｔｍａｘは、減衰曲線ｅ_ｏ（ｔ）_ＦＩＲについて、そのピーク位置の時刻とその振幅が略ゼロになる時刻との時間差である。

For example, the attenuation curve of the impulse response for the acoustic space drawn by superimposing the respective peak positions is defined as e _o (t) _IR, and the attenuation curve of the filter coefficient sequence is defined as e _o (t) _FIR. When the value X calculated according to the above is used as the first index, 450 or a value in the vicinity thereof may be used as the first threshold.

However, the Σ operation of Equation 1 is executed only for the time t satisfying the following Equation 2, and in Equation 1, the time at the peak position and the amplitude of the attenuation curve e _o (t) _FIR are substantially zero. It is the time difference from the time of becoming.

  Here, the reason why 450 or a value in the vicinity thereof is used as the first threshold value to be compared with the first index (hereinafter referred to as the first index X) calculated according to Equation 1 is as follows. According to an experiment conducted by the present applicant, as the first index X calculated according to Equations 1 and 2 exceeds 450, the harsh distortion in the reproduced sound of the speaker (hereinafter referred to as “gearly”). On the other hand, when the first index X is significantly lower than 450, the reflected sound and reverberation sound applied by the filter processing by the FIR filter are buried in the impulse response of the acoustic space, and the FIR filter processing is performed. It became clear that you could not experience the effect of. Therefore, in the aspect in which the first index X is calculated according to Equation 1 and Equation 2, it is desirable to use 450 or a value in the vicinity thereof as a value to be compared with the first indicator X.

On the other hand, a correlation coefficient is an example of an index representing the similarity between two curves. For example, in the aspect in which the correlation coefficient ρ _xy calculated for the attenuation curve of the filter coefficient example drawn by superimposing the respective peak positions and the attenuation curve of the impulse response of the acoustic space 1 is used as the second index. A value of 0.75 or more may be used as the second threshold value to be compared with the second index by the filter coefficient string setting means. Here, the correlation coefficient ρ _xy is calculated according to the following equations 3 and 4 with respect to the attenuation curve e _o (t) _IR of the impulse response in the acoustic space and the attenuation curve e _o (t) _FIR of the filter coefficient sequence. According to the experiment conducted by the present applicant, when the value of the correlation coefficient ρ _xy is 0.75 or more, the FIR tap is separated from the sound due to the characteristic that the FIR tap is separated in the reproduced sound of the speaker. It turned out that suppression of an unnatural feeling (henceforth "para-palm feeling") is remarkable.

The best mode for carrying out the present invention will be described below with reference to the drawings.
(A: 1st Embodiment)
FIG. 1 is a diagram showing an overall configuration of a sound field support system including a sound field support device 40 according to an embodiment of the present invention. This sound field support system controls the acoustic characteristics in the acoustic space 1. Control of acoustic characteristics in the present embodiment means controlling reflected sound characteristics including characteristics such as the level of initial reflected sound, time structure, and arrival direction, and reverberation characteristics. This sound field support system includes a microphone 10-k (k = 1 to 8), a speaker 20-m (m = 1, 2,... M), a microphone amplifier unit 31, a power amplifier unit 32, and a sound field support device 40. Have. The microphone 10-k (k = 1 to 8) and the speaker 20-m (m = 1, 2,... M) are fixed to the side wall and ceiling of the acoustic space 1 with a space therebetween.

  In the sound field support device 40, sound generated by a musical instrument or the like in the acoustic space 1 is converted into a microphone 10-k (k = 1 to 8), the sound field support device 40, and a speaker 20-m (m = 1, 2). ... a state of returning to the acoustic space 1 via M) (referred to as an “acoustic feedback state”), and when the sound passes through the sound field support device 40, the sound related to the reverberation of the acoustic space 1 and the early reflection sound Signal processing is performed so as to approximate the acoustic characteristics of another acoustic space whose characteristics are targeted (referred to as “target acoustic space”).

  FIG. 2 is a diagram illustrating a configuration of the sound field support device 40. In this sound field support device 40, an 8-channel analog signal input from the microphone 10-k (k = 1 to 8) that picks up the sound through the microphone amplifier unit 31 is converted into an A / D converter (not shown). ), And after passing through a mixer 51 and an EMR (Electronic Microphone Rotator) 52, input to a FIR (Finite impulse response) filter 53-i (i = 1 to 4) as four collected sound signals. Is done. Here, the EMR 52 functions to flatten the frequency characteristics of the system by electrically switching the connection relation between the four systems of signals input to the EMR 52 and the four systems of signals output from the EMR 52 from time to time. Fulfill.

The FIR filter 53-i (i = 1 to 4) is a signal obtained by delaying the sound pickup signal input from the EMR 52 by a delay time t _j (j = 1, 2,... G) (“delayed audio signal s _j (j = 1, 2,... G) ”, and delays filter _coefficient values h _j (j = 1, 2,... G) corresponding to each of these delay times t _j (j = 1, 2,... G). Multiplying the audio signals s _j (j = 1, 2,... G) and adding each of the multiplication results is performed, and the product-sum operation result is used as a signal (“reverberation signal”). ").

A reverberation pattern is set in each of the FIR filters 53-i (i = 1 to 4). The reverberation pattern is a pair of filter _coefficient values h _j (j = 1, 2,... G) corresponding to an impulse response (time response) of the target acoustic space and a delay time t _j (j = 1, 2,... G). Filter coefficients ht _j (j = 1, 2,... G) are set, and these filter coefficients ht _j (j = 1, 2,... G) are arranged in time axis order. When the impulse response (time response) in the filter coefficient ht _j (j = 1, 2,... G) is Fourier-transformed, the frequency response of the target acoustic space is obtained. The frequency response obtained by the Fourier transform includes an amplitude characteristic and a phase characteristic in the target acoustic space. FIGS. 3A and 3B show filter coefficient values h _j (j = 1, 2,... G) forming reverberation pattern filter coefficients ht _j (j = 1, 2,... G) and a delay time t _j (j = 1, 2,... G) is an example. As shown in FIG. 3 (A) and FIG. 3 (B), the reverberation pattern set in each of the FIR filters 52-i (i = 1 to 4) has a time as shown in FIG. 3 (A). Various things can be considered, such as those that abruptly attenuate with the passage of time, and those that attenuate relatively slowly as shown in FIG.

  In FIG. 2, the reverberation signal output from the FIR filter 53-i (i = 1 to 4) is equalized by a PEQ (Parametric Equalizer) 54-i (i = 1 to 4) and the compressor 55-i (i = 1). After passing through the dynamic range compression in (4) to (4), the signals are input to the level / delay matrix 58 via the switch 56-i (i = 1 to 4) and the adder 57-i (i = 1 to 4). The switch 56-i (i = 1 to 4) plays a role of switching the acoustic feedback state between on and off. The adder 57-i (i = 1 to 4) plays a role of supplying the output signal to the level / delay matrix 58 when a signal is output from the noise generator 64. As will be described in detail later, the output signal of the noise generator 64 is used when measuring the impulse response of the acoustic space 1.

  The level delay matrix 58 includes four input signal lines (not shown) connected to the adders 57-i (i = 1 to 4) and output signal lines (not shown) for M channels connected to the power amplifier unit 32. ), And a variable resistor for gain adjustment (not shown) and a delay element (not shown) are arranged at each of the crossing positions. The four systems of reverberation signals input to the level delay matrix 58 are subjected to gain adjustment and phase adjustment at the crossing positions of the respective input signal lines and the output signal lines of the respective channels 1 to M, and M channels. It is divided into. Each of the divided M-channel reverberation signals is converted into an analog signal by a D / A converter (not shown), and the converted analog signal is amplified by the power amplifier unit 32 and then the speaker 20-m. (M = 1, 2,... M).

The CPU 61 is a control center of the sound field support apparatus 40 and executes a control program stored in the ROM 63 while using the RAM 62 as a work area. The ROM 63 stores an FIR library (not shown) in which a plurality of types of reverberation patterns are stored. The CPU 61 operating according to the control program can select an acoustic space from the reverberation patterns stored in the FIR library. A process (hereinafter referred to as a filter coefficient sequence optimization process) of selecting one corresponding to the attenuation characteristic of the impulse response of 1 and setting it in each of the FIR filters 53-i is executed. In this embodiment, the FIR library is stored in the ROM 63 in advance. However, the FIR library is stored in an external memory such as a USB memory, and the external memory is stored in the sound field support device via the USB interface or the like. It is also possible to cause the CPU 61 to execute processing for reading a reverberation pattern from the FIR library stored in the external memory. The operation unit 65 in FIG. 2 is an input device including a plurality of operators, and is used for causing the user to perform various input operations such as inputting an instruction to start execution of the program.
Hereinafter, the filter coefficient sequence optimization process, which is a feature of the present embodiment, will be described in detail.

  FIG. 4 is a diagram for explaining the flow of the filter coefficient sequence optimization process. This filter coefficient sequence optimization process is roughly divided into three steps. The first step is a process of acquiring data representing the attenuation curve of the impulse response of the acoustic space 1 and the attenuation curve of the filter coefficient sequence for each of a plurality of types of reverberation patterns stored in the FIR library (FIG. 4: First attenuation curve data acquisition process S01a and second attenuation curve data acquisition process S01b). Either the first attenuation curve data acquisition process S01a or the second attenuation curve data acquisition process S01b may be executed first, or both may be executed in parallel. The second step is a process of calculating data serving as an index for selecting a reverberation pattern set in each of the FIR filters 53-i (i = 1 to 4) from data representing the attenuation curves (FIG. 4). : First index calculation process S02a and second index calculation process S02b). Either of the first index calculation process S02a and the second index calculation process S02b may be executed first, or both may be executed in parallel. The third step selects a filter coefficient sequence (reverberation pattern) for each of the FIR filters 53-i (i = 1 to 4) based on the index calculated in the second step, and the reverberation pattern. Is a filter coefficient string setting process S03.

The first attenuation curve data acquisition process S01a in FIG. 4 is a process for acquiring first attenuation curve data representing the attenuation curve e (t) _IR of the impulse response of the acoustic space 1. If it demonstrates in detail, CPU61 will acquire the data which show the waveform of the impulse response of the acoustic space 1 first in the following ways. First, the CPU 61 switches the switch 56-i (i = 1 to 4) to the OFF state, and the FIR filter 53-i (i = 1 to 4) is input to the FIR filter 53-i. Switch to a state that outputs without going through ("bypass through state"). Next, the CPU 61 causes the noise generator 64 to generate an impulse response measurement signal (for example, a pink noise signal) for a predetermined time (for example, 2 seconds). The signal output from the noise generator 64 in this way is supplied to the CPU 61, and also via the adder 57-i (i = 1 to 4), the level / delay matrix 58 and the power amplifier unit 32. -M (m = 1 to M) is radiated to the acoustic space 1 as a test sound. The response sound of the test sound in the acoustic space 1 is collected by the microphone 10-k (k = 1 to 8), and the collected sound signal is input to the CPU 61 via the microphone amplifier unit 31, the mixer 51, and the EMR 52. . The CPU 61 performs the calculation shown in the following equation 5 on the sampling data g (s) representing the sound collection signal, thereby the first attenuation curve data representing the attenuation curve e (t) _IR of the impulse response of the acoustic space 1. To get. In this embodiment, the calculation shown in Equation 5 is performed with τ = 25 milliseconds. However, what value should be used for τ may be determined by performing experiments as appropriate.

The second attenuation curve data acquisition process S01b in FIG. 4 acquires the second attenuation curve data representing the attenuation curve e (t) _FIR of the filter coefficient sequence in each of a plurality of types of reverberation patterns stored in the FIR library. It is processing to do. More specifically, the CPU 61 performs the operation shown in Equation 5 above (more specifically, in Equation 5) with respect to the combination of the filter coefficient value and the delay time in each of a plurality of types of reverberation patterns stored in the FIR library. A second attenuation curve representing the attenuation curve e (t) _FIR of the impulse response by performing an operation in which the integration is replaced with the sum for the delay time t _j and g (s) is replaced with the filter _coefficient value ht _j. Get the data.
The first and second attenuation curve data obtained as described above (in FIG. 4, these data are expressed as e (t) _IR and e (t) _FIR ) are the first index calculation process S02a and This is input data for the second index calculation process S02b.

The first index calculation process S02a in FIG. 4 is a process for calculating an index related to the crispness of the reproduced sound of the speaker from the first and second attenuation curve data. It is conceivable to use various indicators as the index indicating such a crispness. In this embodiment, the value obtained by performing the calculation shown in the following equation 7 for the time t satisfying the following equation 6: X is used as the first index. In Equations 6 and 7, each of e _o (t) _IR and e _o (t) _FIR represents an attenuation curve e (t) _IR of the impulse response of the acoustic space 1 and an attenuation curve e (t) of the filter coefficient example. It is an attenuation curve obtained by applying an offset in the time axis direction so that peak positions overlap each other in each _FIR . For example, the attenuation curve e (t) _IR of the impulse response of the acoustic space 1 is the waveform shown in FIG. 5A, and the attenuation curve e (t) _{FIR of the} filter coefficient sequence is the waveform shown in FIG. In this case, as shown in FIG. 5C, the CPU 61 applies an offset in the time axis direction to t _i ′ = t _{i + 2} (i = 0, 1, 2,...) For e (t) _IR. e _o (t) _IR is obtained, and for the e (t) _FIR , the attenuation curve e _o (t is obtained by applying an offset in the time axis direction as t _i ′ = t _{i + 4} (i = 0, 1, 2,...). ) Get the _FIR . Further, the CPU 61 calculates a time difference tmax (see FIG. 5C) between the time at the peak position and the time at which the amplitude becomes substantially zero for the attenuation curve e _o (t) _FIR obtained as described above. To do. This tmax is used when the calculation shown in Equation 7 is performed.

数８においてｍおよびｎは任意の自然数（ただし、ｍ＝１のときはｎ≠１、ｎ＝１のときはｍ≠１） X is calculated according to Equation 7, the hatched at attenuation curve _e o _(t) decay curve _e o _{(t) FIR} portion of the filter coefficient string protruding from _IR (FIG. 6 of the impulse response of the acoustic space 1 It is a value obtained by multiplying the area (shown part) by the weight (t / tmax in Equation 7) corresponding to the position of each part on the time axis (that is, time t). The reason why the value X obtained by the calculation shown in Equation 7 is used as an index for evaluating the crispness is that the larger the area of the protruding portion and the closer the portion is to tmax, the larger the speaker. This is because the experimental result that the crispness in the reproduced sound becomes stronger is obtained. In the present embodiment, the value X calculated according to Equation 7 is used as an index representing the crispness of the reproduced sound of the speaker. However, the value X ′ obtained by the calculation shown in Equation 8 below may be used. The point is that when the attenuation curve of the filter coefficient sequence and the attenuation curve of the impulse response of the acoustic space 1 are drawn with the peak positions overlapped, the respective ones of the former attenuation curves protruding from the latter attenuation curve. Any value obtained by multiplying the degree of protrusion by the weight corresponding to the position on the time axis occupied by those portions may be used.

In Equation 8, m and n are arbitrary natural numbers (however, n ≠ 1 when m = 1, m ≠ 1 when n = 1)

The second index calculation process S02b in FIG. 4 is a process for calculating an index related to the feeling of parallax of the speaker playback sound from the first and second attenuation curve data. It is conceivable to use various indicators as such a feeling of parallax, but in this embodiment, the value ρ _xy obtained by the calculation shown in the following equation 9 is used. In Equation 9, σ _x is a standard deviation for e _o (t) _FIR , and σ _y is a standard deviation for e _o (t) _IR . In addition, σ _{xy in Equation} 9 is a value calculated according to Equation 10 below, and <e _o (t) _FIR > in _Equation 10 is an average value of e _o (t) _FIR , <e _o (t) _IR > is the average of e _o (t) _IR . That is, the value [rho _xy is calculated according to Equation 9, the correlation coefficient of the attenuation curve _e o and _{(t) FIR} decay curve _e o _{(t) IR,} represents the similarity of both decay curves.

Here, the value indicating the similarity between the waveform of the attenuation curve e _o (t) _{FIR and} the waveform of the attenuation curve e _o (t) _IR is used as an index for selecting a filter coefficient sequence for the following reason. FIG. 7 is a graph depicting an attenuation curve e _o (t) _IR and an attenuation curve e _o (t) _{FIR of} a certain filter coefficient sequence. The waveforms of both are large in a portion surrounded by a dotted line in FIG. Is different. Experiments conducted by the present applicant have revealed that this difference in waveform is perceived as the above-mentioned feeling of parallax. In other words, in order to suppress the parallax in the reproduced sound of the speaker, a filter coefficient sequence having an attenuation curve having a high similarity to the attenuation curve of the impulse response in the acoustic space 1 is used as the FIR filter 53-i (i = 1 to 4). Should be set for each of the above. This is the reason why a value indicating the similarity between the waveform of the attenuation curve e _o (t) _{FIR and} the waveform of the attenuation curve e _o (t) _IR is used as an index for selecting a filter coefficient sequence. In the present embodiment, the correlation coefficient is used as an index indicating the similarity between the waveform of the attenuation curve e _o (t) _{FIR and} the waveform of the attenuation curve e _o (t) _IR. good. As an example of such an index, for example, when the attenuation curve e _o (t) _FIR and the attenuation curve e _o (t) _IR are expressed as coordinate points in a multidimensional space (for example, Euclidean It is conceivable to use distance: the shorter the distance between these two points, the more similar the attenuation curves.

The filter coefficient string setting process S03 in FIG. 4 includes the first index (X in the present embodiment) calculated by the first index calculation process S02a and the second index calculated by the second index calculation process S02b. Based on (in this embodiment, ρ _xy ), one to be set for each of the FIR filters 53-i (i = 1 to 4) is selected from a plurality of types of reverberation patterns stored in the FIR library. In this process, the setting is performed. More specifically, in the filter coefficient sequence setting process S03, the CPU 61 determines in advance a first index calculated for the attenuation curve of the filter coefficient sequence from among a plurality of types of reverberation patterns stored in the FIR library. A value that is a value in the vicinity of the first threshold value and whose similarity indicated by the second index is equal to or greater than a predetermined second threshold value is selected and set.

Here, as for the first threshold value for performing the evaluation on the first index and the second threshold value for performing the evaluation on the second index, what are the first and second indices? It is necessary to determine appropriately according to whether the type is used. However, it should be noted that the first threshold value is not too small. When the too small the first threshold value, the filter coefficient sequence setting processing S03, for example, the attenuation curve _e o _(t) most of the _FIR attenuation curve _e o _(t) filter coefficient sequence as below _IR (i.e., FIR A filter coefficient sequence in which the frequency characteristics of the filter are buried in the frequency characteristics of the acoustic space) is set in the FIR filter 53-i, and the FIR filter 53-i does not perform its original function of generating reverberant sound or reflected sound. This is because there is a fear. For example, when X calculated according to the above-mentioned number 7 is used as the first index, it is preferable to use 450 or a value in the vicinity thereof as the first threshold, and to the above-mentioned number 9 as the second index. Therefore, when using the calculated ρ _xy , it has been found by experiments conducted by the present applicant that it is preferable to use a value of 0.75 or more as the second threshold value.

  Various modes are conceivable as modes for determining the filter coefficient sequence to be set in the FIR filter 53-i based on the first and second indices, and examples thereof are as follows. First, among filter coefficient sequences in which the value of the second index is greater than or equal to a predetermined second threshold value, the one having the value of the first index closest to the predetermined first threshold value is FIR. The mode set in the filter 53-i and the similarity represented by the second index among the filter coefficient sequences in which the value of the first index falls within a predetermined range centered on a predetermined first threshold value Is a mode in which the maximum is set in the FIR filter 53-i. According to these aspects, it is possible to automatically set the filter coefficient sequence suitable for the acoustic characteristics of the acoustic space 1 in the FIR filter 53-i without causing the user to perform a special operation.

  The second aspect is a predetermined range in which the value of the first index is centered on the first threshold (for example, when the first threshold is 450, a range of ± 50 centered on this value, etc.) ) And a filter coefficient string whose similarity represented by the second index is equal to or higher than the second threshold value, an identifier for uniquely identifying the filter coefficient string (for example, a plurality of types stored in the FIR library) A series number assigned to each of the reverberation patterns in FIG. 8, the first and second indices calculated with respect to the filter coefficient sequence, and tmax in the manner shown in FIG. 8 on a display device (not shown). In this mode, the user selects a filter coefficient sequence to be set in each of the FIR filters 53-i (i = 1 to 4) by an operation on the operation unit 65. According to this aspect, it is possible to flexibly select the filter coefficient sequence to be set in the FIR filter 53-i, for example, by selecting the one having the largest value of tmax (that is, the one having the longest reverberation time). Even when the filter coefficient string is selected based on the value of tmax as described above, the filter coefficient string displayed in the list satisfies both of the first and second indices. Therefore, it is needless to say that the crispness and paralysis of the reproduced sound of the speaker when the filter coefficient sequence is set to the FIR filter 53-i is suppressed.

  As described above, according to the present embodiment, in the sound field support system, not only the occurrence of problems such as howling and sound coloration can be avoided, but also an acoustic engineer can reproduce each of the speakers 20-m. There is an effect that it is possible to improve the sound quality of the reproduced sound without actually listening to and confirming the sound.

(B: Other embodiments)
Although one embodiment of the present invention has been described above, the following embodiments are also conceivable.
(1) In the above-described embodiment, the impulse response for the acoustic space 1 is measured, and attenuation curve data representing the attenuation curve is calculated and acquired from the measurement result. However, an external memory storing data indicating the measurement result of the impulse response of the acoustic space 1 is connected to the sound field support device 40, and attenuation curve data representing the attenuation curve is calculated from the data stored in the external memory. You may do it. Further, an external memory storing attenuation curve data representing an attenuation response curve of the acoustic space 1 may be connected to the sound field support device 40, and the attenuation curve data stored in the external memory may be read. Similarly, data representing the attenuation curve of each filter coefficient sequence may be calculated in advance, and the filter coefficient value / delay time pair and the data representing the attenuation curve may be used as a reverberation pattern.

(2) In the embodiment described above, a control program for causing the CPU 61 to execute the filter coefficient sequence optimization process (see FIG. 4) characteristic of the present invention is stored in the ROM 63 in advance. However, the program may be distributed by being written on a computer-readable recording medium such as a CD-ROM (Compact Disk-Read Only Memory), or the program may be downloaded via a telecommunication line such as the Internet. You may distribute it. The programs distributed as described above are stored in a general computer device (for example, a personal computer), and the control unit (CPU) of the computer device is operated according to the program, thereby the general computer. This is because the same function as the sound field support device 40 according to the above embodiment can be given to the device.

(3) In the above-described embodiment, the filter coefficient sequence optimization process (FIG. 4) characteristic of the present invention is realized by software, but may of course be realized by hardware. Specifically, the first attenuation curve data representing the attenuation curve of the impulse response of the acoustic space (for example, the acoustic space 1 in the above embodiment) to be subjected to sound field control, and the attenuation curve of each filter coefficient sequence are represented. Attenuation curve data acquisition means for acquiring second attenuation curve data, calculation means for calculating data as an index for selecting a filter coefficient string, selection of a filter coefficient string based on the index, and setting to the FIR filter Each of the filter coefficient string setting means for performing the above may be configured by an electronic circuit, and the sound field support apparatus may be configured by combining these means.

It is a figure which shows the structural example of the sound field assistance system containing the sound field assistance apparatus 40 which concerns on one Embodiment of this invention. It is a figure which shows the structural example of the same sound field assistance apparatus. It is a figure which shows an example of the reverberation pattern set to FIR filter 53-i (i = 1-4) of the same sound field assistance apparatus 40. FIG. It is a figure which shows the flow of the filter coefficient sequence optimization process which CPU61 of the same sound field assistance apparatus 40 performs. It is a figure for demonstrating the offset provision process which the CPU61 performs by the 1st and 2nd parameter | index calculation process. It is a figure for demonstrating the meaning of the 1st parameter | index X calculated by 1st parameter | index calculation process S02a included in the filter coefficient sequence optimization process. It is a figure for demonstrating the meaning of 2nd parameter | index (rho) _xy calculated by 2nd parameter | index calculation process S02b included in the same filter coefficient sequence optimization process. It is a figure which shows an example of UI screen for making a user select the filter factor sequence set to a FIR filter from the setting candidates narrowed down based on the 1st parameter | index X and the 2nd parameter | index (rho) _xy . It is a figure for demonstrating operation | movement of a sound field assistance system.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Acoustic space, 10-k (k = 1-8) ... Microphone, 20-m (m = 1-M) ... Speaker, 31 ... Microphone amplifier part, 32 ... Power amplifier part, 40 ... Sound field assistance apparatus, 51 ... Mixer, 52 ... EMR, 53-i (i = 1-4) ... FIR filter, 54-i (i = 1-4) ... PEQ, 55-i (i = 1-4) ... Compressor, 56- i (i = 1 to 4) ... switch, 57-i (i = 1 to 4) ... adder, 58 ... level delay matrix, 61 ... CPU, 62 ... RAM, 63 ... ROM, 64 ... noise generator, 65 ... operation part.

Claims (5)

While acquiring the 1st data showing the decay curve of the impulse response of acoustic space, it connects with the speaker arrange | positioned in the said acoustic space, and the microphone which picks up the sound output from the said speaker, and forms a closed loop. A second filter coefficient sequence is selected from an FIR library storing a plurality of types of filter coefficient sequences respectively set in an FIR (Finite Impulse Response) filter, and a second curve representing an attenuation curve for the filter coefficient sequence is selected. Attenuation curve data acquisition means for acquiring the data of
For each of the second data, the former protruding from the latter attenuation curve when the attenuation curve represented by the second data and the attenuation curve represented by the first data are drawn with the peak positions superimposed. A first index obtained by multiplying the degree of protrusion of each part of the attenuation curve by a weight corresponding to the position on the time axis occupied by the part, and an attenuation curve represented by the first data A second index indicating the degree of similarity; an index calculating means for calculating from the first and second data;
Of the plurality of types of filter coefficient sequences, the first index is a value in the vicinity of a predetermined first threshold value, and the similarity indicated by the second index is a predetermined second value. And a filter coefficient sequence setting unit that selects and sets the FIR filter that is equal to or greater than a threshold value. The first index is an attenuation curve e _o (t) _IR of the impulse response of the acoustic space drawn by superimposing each peak position and an attenuation curve e _o (t) _{FIR of the} filter coefficient sequence. The sound field support apparatus according to claim 1, wherein the sound field support apparatus is a value X calculated according to the following formula 1 and the first threshold value is 450 or a value in the vicinity thereof.

However, the Σ operation in Equation 1 is executed only for time t that satisfies Equation 2 below, and tmax in Equation 1 is substantially zero at the time of its peak position and its amplitude for the attenuation curve e _o (t) _FIR. It is the time difference from the time.

  The second index is a correlation coefficient calculated with respect to the attenuation curve of the impulse response of the acoustic space and the attenuation curve of the filter coefficient sequence drawn by superimposing the respective peak positions, The sound field support apparatus according to claim 1 or 2, wherein the threshold value is 0.75 or more. While acquiring the 1st data showing the decay curve of the impulse response of acoustic space, it connects with the speaker arrange | positioned in the said acoustic space, and the microphone which collects the sound output from the said speaker, and forms a closed loop. A second filter coefficient sequence is selected from an FIR library storing a plurality of types of filter coefficient sequences respectively set in an FIR (Finite Impulse Response) filter, and a second curve representing an attenuation curve for the filter coefficient sequence is selected. A first step of obtaining the data of
For each of the second data, the former protruding from the latter attenuation curve when the attenuation curve represented by the second data and the attenuation curve represented by the first data are drawn with the peak positions superimposed. A first index obtained by multiplying the degree of protrusion of each part of the attenuation curve by a weight corresponding to the position on the time axis occupied by the part, and an attenuation curve represented by the first data A second step of calculating a second index indicating similarity from the first and second data;
Of the plurality of types of filter coefficient sequences, the first index is a value in the vicinity of a predetermined first threshold value, and the similarity indicated by the second index is a predetermined second value. A sound field support method, comprising: a third step of selecting one that is equal to or greater than a threshold and setting the FIR filter. Computer equipment,
While acquiring the 1st data showing the decay curve of the impulse response of acoustic space, it connects with the speaker arrange | positioned in the said acoustic space, and the microphone which collects the sound output from the said speaker, and forms a closed loop. A second filter coefficient sequence is selected from an FIR library storing a plurality of types of filter coefficient sequences respectively set in an FIR (Finite Impulse Response) filter, and a second curve representing an attenuation curve for the filter coefficient sequence is selected. Attenuation curve data acquisition means for acquiring the data of
For each of the second data, the former protruding from the latter attenuation curve when the attenuation curve represented by the second data and the attenuation curve represented by the first data are drawn with the peak positions superimposed. A first index obtained by multiplying the degree of protrusion of each part of the attenuation curve by a weight corresponding to the position on the time axis occupied by the part, and an attenuation curve represented by the first data A second index indicating the degree of similarity; an index calculating means for calculating from the first and second data;
Of the plurality of types of filter coefficient sequences, the first index is a value in the vicinity of a predetermined first threshold value, and the similarity indicated by the second index is a predetermined second value. A program that functions as a filter coefficient string setting unit that selects a value that is equal to or greater than a threshold value and sets it in the FIR filter.

Images